Critical gaps across all 18 tools. No visible input schemas in provided source code, cannot verify parameter types, constraints, or descriptions. Tool descriptions are present but generic (10-50 chars), lacking WHAT/WHEN/WHY context required by LLMs. No documented output schemas. Tool names lack action verbs (e.g., 'chat', 'clink', 'thinkdeep' are vague; should be 'start_chat', 'bridge_to_cli', 'analyze_deeply'). No evidence of error handling, recovery guidance, or parameter validation. Base infrastructure (BaseTool) exists but actual tool implementations not visible in source excerpt.

• Add explicit JSON Schema input definitions for all 18 tools with type, description, and constraints for every parameter. Example: { 'type': 'object', 'properties': { 'code': { 'type': 'string', 'description': 'Source code to analyze (max 50KB)' }, 'language': { 'type': 'string', 'enum': ['python', 'javascript', 'go'], 'description': 'Programming language' } }, 'required': ['code'] }
• Rename tools to start with action verbs: 'chat' → 'start_interactive_chat', 'clink' → 'bridge_to_ai_cli', 'thinkdeep' → 'perform_deep_analysis', 'analyze' → 'analyze_code', 'lookup' → 'search_knowledge_base'. Verify names match verb_noun pattern.
• Expand all tool descriptions to 50-200 chars, answering: WHAT does it do? WHEN should LLM call it? WHAT does it return? Example: 'Perform step-by-step code review with expert analysis. Call when user requests code feedback. Returns structured review with issues, severity, and fix suggestions.'
• Document output schemas for all tools. Specify return type (object/array), required fields, and field types. Include pagination (limit, offset, total_count) for list tools. Example: { 'type': 'object', 'properties': { 'issues': { 'type': 'array', 'items': { 'type': 'object', 'properties': { 'line': { 'type': 'integer' }, 'message': { 'type': 'string' }, 'severity': { 'enum': ['error', 'warning', 'info'] } } } } }
• Add error handling with recovery guidance. Return structured errors: { 'error': 'invalid_language', 'message': 'Language must be one of: python, javascript, go. Got: rust.', 'recovery': 'Call analyze with a supported language.' }. Categorize as retryable, user-fixable, or fatal.

Inferred effective spec: 2026-07-28+.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry